MVTec Software GmbH

Sample-based Identification (SBI)

Featured in HALCON
Sample-based Identification (SBI)
The data set consists of 14 different vegetables and fruits. For each object group only one training sample was used.

Identification based only on the visual features of an object is now possible thanks to a new technology from MVTec. With minimal training, MVTec’s sample-based identification (SBI) technology can very quickly identify a trained object without the need for special codes or markings.

Additionally, SBI can rapidly differentiate between a large set of trained objects. With applications ranging from mobile marketing technology to grocery retail,  SBI utilizes characteristic visual features like color and texture to properly identify a trained object, thereby eliminating the need to use special imprints like bar codes or data codes.

 

SBI works with warped objects or varying perspective views of the object. Training can be accomplished with as little as a single image, or for more robust performance, it’s possible to  use multiple images showing different views of  the object.  This allows 3D objects to be robustly recognized from any viewing angle.